Draft: lowered cuts in the seeding to allow for cosmics
The goal is to get the cosmics that still are roughly parallel to the beam line and pass through the UT. To be honest, I am unsure this will be successful: a cosmics would still need to be in a very narrow window in terms of geometry and timing.
Things will be refined further, this is a very fast, untested attempt, to showcase what I think the changes should be.
Rationale behind the changes
Three cuts have been modified, the two-hit cut, three-hit cut, and y tolerance.
Two-hit Cut:
The idea behind the cut is that a particle coming from (0,0,0) with infinite momentum is projected around a certain point if a first measurement is known. Now, our cosmics actually passes through (x0Prime,0,0) (I have conflated xMaxClosestToBeam with xAtZ0 to keep the number of parameters small), so we correct the tolerance window accordingly
Three-hit Cut:
The idea behind the current cut is that a particle from (0,0) that passes through (x1,z1) and (x2,z2) has seen its slope change from x1/z1 to (x2-x1)/(z2-z1), and we can use this to calculate the momentum. We introduce a term to inflate the 1/p estimation according to the maxXClosestToBeam.
y Cut:
The most difficult to adapt since we actually should change the logic behind it. For the first-hit collection, I inflate the tolerance to illustrate how little we know about the actual slope due to our uncertainty on the origin region. Once the first hit is found (for now, always first layer considered, must be changed), we replace rather than update our ty assumption, and then continue as normal. It seems there are no cuts on the ay parameter, need to be careful about subsequent cuts: normal seeding running assumes a very small (technically, 0) value, but we should here allow up to 750 mm.